Blab Tech
JPMorgan Chase to Deploy Long-Running AI Agents This Year
By the end of 2026, JPMorgan Chase plans to roll out a new generation of artificial‑intelligence agents that can operate autonomously for hours, a move that signals the bank’s readiness to overcome the security and governance hurdles that have slowed large‑company adoption of generative AI.
The announcement followed an interview with JPMorgan’s chief analytics officer, Derek Waldron, on CNBC. Waldron explained that the firm is shifting from short‑lived chat‑bot interactions to “long‑running autonomous agents” that can maintain context and complete multi‑step workflows across different software systems for extended periods, rather than stopping after a few minutes of instruction.
Long‑running agents already exist in the public domain. Anthropic’s Claude Code, a coding assistant that can edit files and run commands, and OpenClaw, an open‑source agent that can automate tasks via messaging platforms, have demonstrated the ability to perform tasks for longer than traditional LLM‑based tools. JPMorgan’s planned deployment suggests the technology is close to meeting the bank’s stringent security and governance requirements.
JPMorgan, the largest U.S. bank by assets and a systemically important institution, operates a technology budget of nearly $20 billion. CEO Jamie Dimon has overseen the bank’s expansion into technology‑driven services for more than a decade, and the firm’s focus on AI has shifted from cost reduction to revenue enhancement, according to Waldron.
Waldron said the key metric for AI agents is “intellectual coherence” – how long a system can operate effectively before needing human intervention. Advances in reasoning and task decomposition allow agents to act as team managers, delegating subtasks and coordinating across multiple applications. The agents can write code, control web browsers, and interact with desktop software, enabling them to handle complex, multi‑step processes.
Security concerns remain a barrier to immediate deployment. Waldron noted that the bank expects fully operational long‑running agents by 2026 and anticipates that future agents could remain coherent for hours, days, or even weeks.
The bank has already seen productivity gains from AI in back‑office functions and software development. In private banking, AI systems now scan market activity, client positions, and research overnight, freeing bankers to focus on client interactions. According to the bank, these tools have contributed to a 20 % increase in gross sales and could allow individual bankers to expand client coverage by up to 50 %.
Dimon has acknowledged that some employees will be displaced by AI, but the bank is preparing to retrain and redeploy affected staff. Waldron noted that many enterprises initially viewed AI as a cost‑cutting tool, but the current trend is toward using AI to create sustainable competitive advantage.
The shift also affects JPMorgan’s approach to software acquisition. The firm is evaluating whether to build capabilities in-house rather than buying from external vendors, a strategy that could pressure traditional software providers. Waldron said the moat around certain types of software companies has narrowed.
JPMorgan’s move reflects a broader industry trend toward autonomous AI agents that can manage end‑to‑end workflows. As banks and other large organizations prepare to deploy these systems, they will need to address challenges related to data privacy, regulatory compliance, and the potential for unintended consequences.
In summary, JPMorgan Chase is set to introduce long‑running AI agents later this year, marking a significant step toward autonomous digital workers that can operate for extended periods. The bank’s experience in private banking and its focus on revenue generation suggest that these agents will play a key role in enhancing client services and operational efficiency. While security and governance hurdles remain, the bank’s timeline indicates that the technology is close to meeting the rigorous standards required for corporate deployment.
The announcement followed an interview with JPMorgan’s chief analytics officer, Derek Waldron, on CNBC. Waldron explained that the firm is shifting from short‑lived chat‑bot interactions to “long‑running autonomous agents” that can maintain context and complete multi‑step workflows across different software systems for extended periods, rather than stopping after a few minutes of instruction.
Long‑running agents already exist in the public domain. Anthropic’s Claude Code, a coding assistant that can edit files and run commands, and OpenClaw, an open‑source agent that can automate tasks via messaging platforms, have demonstrated the ability to perform tasks for longer than traditional LLM‑based tools. JPMorgan’s planned deployment suggests the technology is close to meeting the bank’s stringent security and governance requirements.
JPMorgan, the largest U.S. bank by assets and a systemically important institution, operates a technology budget of nearly $20 billion. CEO Jamie Dimon has overseen the bank’s expansion into technology‑driven services for more than a decade, and the firm’s focus on AI has shifted from cost reduction to revenue enhancement, according to Waldron.
Waldron said the key metric for AI agents is “intellectual coherence” – how long a system can operate effectively before needing human intervention. Advances in reasoning and task decomposition allow agents to act as team managers, delegating subtasks and coordinating across multiple applications. The agents can write code, control web browsers, and interact with desktop software, enabling them to handle complex, multi‑step processes.
Security concerns remain a barrier to immediate deployment. Waldron noted that the bank expects fully operational long‑running agents by 2026 and anticipates that future agents could remain coherent for hours, days, or even weeks.
The bank has already seen productivity gains from AI in back‑office functions and software development. In private banking, AI systems now scan market activity, client positions, and research overnight, freeing bankers to focus on client interactions. According to the bank, these tools have contributed to a 20 % increase in gross sales and could allow individual bankers to expand client coverage by up to 50 %.
Dimon has acknowledged that some employees will be displaced by AI, but the bank is preparing to retrain and redeploy affected staff. Waldron noted that many enterprises initially viewed AI as a cost‑cutting tool, but the current trend is toward using AI to create sustainable competitive advantage.
The shift also affects JPMorgan’s approach to software acquisition. The firm is evaluating whether to build capabilities in-house rather than buying from external vendors, a strategy that could pressure traditional software providers. Waldron said the moat around certain types of software companies has narrowed.
JPMorgan’s move reflects a broader industry trend toward autonomous AI agents that can manage end‑to‑end workflows. As banks and other large organizations prepare to deploy these systems, they will need to address challenges related to data privacy, regulatory compliance, and the potential for unintended consequences.
In summary, JPMorgan Chase is set to introduce long‑running AI agents later this year, marking a significant step toward autonomous digital workers that can operate for extended periods. The bank’s experience in private banking and its focus on revenue generation suggest that these agents will play a key role in enhancing client services and operational efficiency. While security and governance hurdles remain, the bank’s timeline indicates that the technology is close to meeting the rigorous standards required for corporate deployment.
